Amazon Can Be Hard to Understand but Easy to Like

There was a fashion, once, for industrial conglomerates.  Companies would buy each other up and the market would reward the result with a higher PE than the components.  This got going in the 1960s and worked into the '70s, but went into reverse with the fashion for down-sizing, right-sizing, and "unlocking value."  A few companies stuck with it as a model, most obviously General Electric.  But the stigma of "too complex to invest in" has lingered over the category ever since.  And most professional investors tell most company managers that, in effect, "we don't need you to diversity, we can do that ourselves in the market."  The mantra for successful company management (and board-level governance) is "stick to your knitting."

It's fair to ask, of some companies, exactly what they think they are knitting here.  Amazon is a case in point.  On the one hand, we think of them as the world's biggest online department store -- more departments selling more different things than anyone else.  They are executing this strategy with a Walmart-like focus on wringing economies of scale.  They're also reshaping the market, for tangible goods anyway, as evidenced in their "Frustration-free Packaging" initiative, which doubles as an eco-minded conservation initiative, and could have a big and welcome impact up the supply chain.  Admirable, and successful, and all consistent with the big-online-store concept.

That very consistency makes it hard to explain Amazon Web Services, a.k.a. "The Cloud."  Now it's true that "cloud computing" is buzzword du jour, but Amazon isn't usually prone to chase fads.  And they've been chasing Web Services for a while -- I recommend the blog by Werner Vogels, their CTO, for the most lucid and hype-free perspective on this.

Here's what I think is really happening, and it's kind of interesting.  The original logic of conglomeration was that management skill was transferable across industries.  Maybe this was vanity, at least for some companies, but clearly GE is on to something with their management training investment and rigorous financial reporting.  Those may well be transferable skills -- competitive advantages that can be honed in some businesses and then applied in others.  What Amazon and Google and Salesforce and Facebook a few others need to have (and think they have) is the ability to operate ultra-large-scale ("cloud") computing facilities reliably and profitably.  That's a super-valuable and highly-specialized skill.  Cloud computing, by this theory, is one of those essential, yet hard to create, skills that will identify the next generation of super-successful global companies.